One of the most commonly used data sources in quant rating is the stock price and volume. We will refer to this as price-volume, but you may also hear it referred to as technical factors. The raw data for price-volume can include anything that we learn from a quotes and trade feed. For example, adjusted and unadjusted prices, the open, high, low, close for each time period. Different time frequencies such as day, week, hour, minute, and bid-ask price quotes. There are operations that we can apply it to the price volume data in which we also defined the window length, which is how many days worth of data to use in the calculation. We’ve already seen returns, which can be daily, weekly, monthly, or yearly. Returns can also be calculated at shorter time scales such as seconds, minutes, hours. Returns can be calculated from the close of one day to the close of the next day. Returns can also be calculated on the open price to close price for the same day. For example, the open price in the morning to the close price of the afternoon on the same day. Returns can be calculated on the close price of one day to the open price of the next day. For example, the return from the close price of Monday evening to the open price of Tuesday morning. The last example is referred to as overnight returns, and we’ll discuss this in detail in a later lesson. We can also apply operations on a distribution of returns. For instance, we can calculate the first four moments, which are called the mean, variance, skew, and kurtosis. The mean describes the center of the return distribution. The variance describes the spread of the distribution. The square root of variance, which is the standard deviation, is a common measure of return volatility. Skew describes the amount of asymmetry. A positive skew, means there are more extreme values in the positive side of the distribution. Kurtosis describes how much of the distribution occurs in the left and right tails. Stock returns distributions tend to exhibit larger tails compared to a normal distribution. Kurtosis is often referred to as fat tails. There’s also the minimum or maximum over a certain time window. Note that we’ll discuss skewness in a later lesson about alpha factors. So, to summarize, pretty much any statistical or time series calculation you can think of, is fair game for use in a price-volume driven factor. Using price and volume as a factor, has a benefit of having readily available data, since stocks trade on exchanges and many market data vendors sell trades, quotes, and bar data. The availability and regularity of the data source is nice to have, since other sources for factor generation such as fundamentals are updated less frequently, and still other data sources such as news, social media, or analyst reports may not always be available for the stock that we’re analyzing. It’s very helpful to have a factor that can generate a signal for every stock in the stock universe. This is especially helpful in quant investing, since quant investing usually involves portfolios of hundreds or thousands of stocks. Keep in mind that the higher frequency of price volume data often leads to more trading, and therefore higher portfolio turnover. Since trading decisions are based on the data, the more often the data updates and changes, the more often the portfolio is likely to require rebalancing. This can be good or bad and requires careful study. Faster signals means more information. More information, all other things being equal, is good. However, more rebalancing means more trades and more transaction costs. So, strategy that is based on frequently updated data generally has more transaction costs. As a researcher, a key thing you need to determine is if the higher information content of the signal, more than offsets, the likely higher transaction costs.